I'm trying to create a spam filter in Java using the Bayesian algorithm. I use a text file that contains email messages and split the tokens using regex, storing these values into a hashmap. My problem is, with regex, the email addresses are split so instead of: [email protected] regex causes the token to be: john smith example The same holds true for ip addresses, so for example, instead of: 192.55.34.322 regex splits the tokens to be: 192 55 34 322 So does anybody know of a way that I could read the email messages and store their contents as is?